IMF Blog IMF Blog

Fair and Substantial—Taxing the Financial Sector

We knew we were in for a tough time when the leaders of the Group of Twenty (G-20) asked the IMF to give them our views, at their summit coming up in June 2010, on  “... the range of options countries have adopted or are considering as to how the financial sector could make a fair and substantial contribution toward paying for any burden associated with government interventions to repair the banking system.”

Everyone has strong feelings these days on the taxation of the financial sector. Taxpayers who financed the rescue of the financial sector during the recent crisis want their money back—or at least not to get caught again. Some want to see more of the money coursing through the financial system turned to public use.

The industry worries about new taxes coming on top of the swathe of regulatory reforms that likely lies ahead for them. And some governments in countries whose financial sector weathered the storm pretty well wonder why they should now ask it for cash. Responding to the request from the G-20 leaders puts us in the middle of all these concerns.

Last week the IMF gave an interim report to the G-20 finance ministers focused on the specific question we were asked: what are the options in raising money from the financial sector to pay for the costs of government intervention from which it benefits. That report is confidential, but—you may have noticed—has still managed to attract a lot of attention.  So let me set out how our thinking on this stands.

Our aim is to take a cool look at the issues about which everyone gets so heated. The options we have come up with so far—this interim report will be revised for the June summit—won’t please everyone (or, maybe, anyone), but can, we hope, help move the public debate ahead by bringing some economic and financial analysis to bear. In doing this we looked forward more than back. Revenge is not a good principle for tax design, but averting and preparing for future harm is.

What is to be done?

The challenge is to ensure that financial institutions bear the direct fiscal costs that any future failures or crises will impose—and maybe somewhat more, given all the other costs that bank failure can impose on the economy.

We also need to make these events both less likely to happen and less costly when they do. We think two types of tax can play a role.

A ‘Financial Stability Contribution’—I come to bury Caesar, not to bail him out

One reason the crisis was such a painful mess was that many governments did not have the tools to wind down failing institutions in a quick and orderly manner. All too often their only options, both hugely unpleasant, were to either (1) let a systemic institution fail and bear the chaotic fallout or (2) pump in enough public support to keep it alive, so confirming the prior suspicion that these institutions were indeed too big to fail. Governments lacked a way to ‘resolve’—a new word even for many economists—large failing institutions.

Resolution means equity holders would be wiped out, management replaced, and unsecured creditors take a loss—a ‘haircut’—on their claims. All this should be nasty enough for owners and managers to reduce any problems of ‘moral hazard’ (taking too much risk in the expectation that someone else will bear the costs if things turn out badly). But most countries still don’t have such a mechanism. Financial stability requires creating them.

So where does the idea of a contribution come in? Resolution requires upfront cash, to reduce uncertainty for creditors (and the creditors’ creditors...) by quickly giving some value to their claims. And the industry should pay for this: it is, or should be, a cost of doing business just like paying for deposit insurance, or maintaining their information systems. This is what we call a Financial Stability Contribution (FSC).

It would ensure that the industry does indeed pay a reasonable chunk of these resolution costs before a crisis occurs, with this amount topped up, if needed, by ‘ex post’ charges after disaster strikes (much as the Financial Crisis Responsibility fee proposed in the United States aims to recoup some of the costs of public support).

Costs of the crisis

Incidentally, as a first step, we tried to figure out how much the recent crisis has cost governments in terms of the direct support they provided to the financial sector. The answer is: so far, about 2.7 percent of GDP for the group of advanced G-20 countries. More for some, less for others – including most emerging market countries.

That’s a sizable sum, but the risks during the crisis were even larger, with guarantees and other contingent liabilities averaging around 25 percent of GDP for the advanced G-20. And all that ignores indirect fiscal costs caused by the recession and (to a lesser extent) stimulus measures—which is causing a surge in public debt—and, perhaps most cruelly, of all, a cumulative loss of output of around 27 percent of GDP.

The FSC would start as a simple levy on some balance sheet (and, possibly, off-balance sheet) variables, but then be refined to strengthen the link with each institution’s contribution to systemic risk—giving them some incentive to reduce it. It would be permanent (to keep that beneficial effect at work, at least until regulatory solutions are felt to have done enough) and paid by all financial institutions (because they all benefit from the greater financial stability the resolution mechanism provides).

Whether the revenue from such a charge should be treated just like other tax revenue or instead feed an earmarked fund to help with resolutions is secondary. The fiscal impact is the same (assuming of course, other policies are not affected by whether there is or not an earmarked fund): the government has to sell fewer bonds on the open market, either because it has more tax revenue or because it has a captive customer in the fund. The main argument for a fund is that it could provide more assurance that the agency in charge of resolution has ready access to the resources it needs.

A ‘Financial Activities Tax’

A FAT is just a tax on the sum of the profits and remuneration paid by financial institutions. That sounds simple, and, in essence, it is. But why an extra tax on financial institutions? Here, I’m afraid, things get a bit nerdy. So brace up for what is coming.

Profits plus all remuneration is value added. So a tax of this kind would be a kind of Value-Added Tax or VAT. And that could make sense because current VATs don’t work well for financial services, which are largely VAT-exempt. This means that a FAT of this kind could make the tax treatment of the financial sector more like that other sectors and so help offset a tendency for the financial sector, purely for tax reasons, to be too large—or too fat.

Now suppose that the base included only remuneration above some high level, and only profits above a ‘normal’ rate of return. Then the base of the FAT may not be a bad proxy for taxes on ‘rents’—return in excess of competitive levels—earned in the sector. Some might find taxing that excess fair.

Or one might include only profits above some level well above normal. Taxing away some of these high returns in good times may help correct for any tendency to excessive risk-taking implied by financial institutions not attaching enough weight to outcomes in bad times (whether because of limited liability, or because they think themselves too big to fail).

What about a financial transactions tax?

We also looked at the idea of a general financial transactions tax (FTT)—the last few months have left us in no doubt as to the seriousness of the public support this enjoys. This would be a tax paid every time a share, bond, or other financial instrument is bought or sold, and/or whenever foreign currency is bought or sold.

Our work is not yet complete—this is an interim report, remember—but, while some forms of FTT may be feasible (indeed most G-20 countries already tax some financial transactions), we don’t think this is the best way of meeting the two key objectives set out above. An FTT is not  focused on reducing systemic risk and it isn’t effective at taxing rents in the financial sector—much of the burden may well fall on ordinary consumers.

Moreover, the financial services industry is very good at devising schemes to get around such a tax and (this is also true, to be fair, of the FSC and FAT, but we suspect to a lesser extent).

One way to think about the comparison is that just as a FAT is like a VAT, an FTT is like a turnover tax—and most countries have long found that the VAT is better at raising revenue: in the jargon, more efficient. All this doesn’t mean we rule out an FTT in other contexts—but it is not the most effective way to address the task at hand.

Should everyone do this?

Several countries that did not need to pour large resources into their financial institutions are naturally reluctant to lumber them with more charges. At the same time, financial institutions are so adept at tax and regulatory arbitrage that those countries who do want to act fear they may be undercut by  those who don’t.

But this tension is not as great as it may seem.

If financial history teaches us anything, it is that no one should think themselves immune from failures and crisis. Moreover, if the FSC in particular is properly risk-adjusted, countries with safer systems will simply face a smaller contribution. And importantly, the last thing we want to do is to repress/bury the financial sector by imposing a heavy burden; like the food supply, it means too many good things for economic growth.


What we gave to G-20 ministers was an interim report, and we will be working more on this in the light of their discussion last Friday.

We will continue too to listen to what others tell us. One theme of our work that has already been widely stressed is that any tax initiatives need to be coordinated with regulatory ones—so we have some number-crunching, as well as more tough times, ahead of us. Still, we hope to contribute to the debate on what really matters in all this: how to reduce the risk, and costliness, of future financial failures.